When I was 12 I was with my family in Southern India. While walking around the market place we saw a man in chains - this was my first encounter with a living slave. Since then I have devoted my life to combatting slavery. This tool is the next iteration in the process of freeing the world from this horrible way of life.
What it does
This tool scrapes sex trafficking websites across the internet. The goal is to repackage the data and sell it to industry. This way companies can internally flag to law enforcement when a trafficker or a victim of trafficking is using their system. This will allow companies to keep traffickers out of their systems and actively aid in reduction of the size of the trafficking marketplace. Traffickers heavily leverage current technology to do their work, by using their own marketing techniques against them - leveraging the data they voluntarily post online, we can drive them deeper into the web, reducing their reach, and limiting their profitability.
How we built it
For this project we made use of Python, Flask and a whole bunch of natural language processing.
Challenges we ran into
Getting some of the features built was hard, due to time constraints.
Accomplishments that we're proud of
We have a fully deployed website available at: https://investagator.herokuapp.com and we've developed a fully functional api. This way we can show the value proposition to potential company clients, with a very limited data set and supply lots and lots of data to end users.
What we learned
We learned a ton at this hackathon! The big take aways were:
- there are lot of dependencies in the code base - so we will be seperating out concerns into multiple repos, so folks can get started with the code base quickly.
What's next for investigator
Extend api to work with graphql (with graphene)
finish image search
Add time series analysis
- w/ pyflux
Develop a chatbot
heatmap - data visualization showing where trafficking is happening
area code to lat long so we can figure out where people are coming from.
oracle of bacon style entity resolution w/ graphene